Abstract
Scenario models of a moving subway train can help investigate the influence of different fire locations on smoke propagation characteristics in curved tunnels. To this end, this study adopts the three-dimensional Unsteady Reynolds Average Navier-Stokes equations method and the renormalization group k-ε two-equation turbulence model with buoyancy correction for numerical analysis. The motion of the train is replicated using the slip grid technique. The results indicate that when a fire breaks out on a moving train in tunnels, the piston wind leads the longitudinal movement of the smoke. If a fire erupts in the head or middle car of a moving train, the time of smoke backflow is delayed by 30 s or 17 s, respectively, compared to that for the tail car. The obtained results provide a theoretical basis for reasonably controlling the smoke flow in subway tunnels and reducing casualties in fire accidents.